Package Details: ik-llama.cpp-cuda-git t0002.r382.528cadb0-1

Git Clone URL: https://aur.archlinux.org/ik-llama.cpp-cuda-git.git (read-only, click to copy)
Package Base: ik-llama.cpp-cuda-git
Description: Port of Facebook's LLaMA model in C/C++ (with NVIDIA CUDA optimizations) - fork by ikawrakow
Upstream URL: https://github.com/ikawrakow/ik_llama.cpp
Licenses: MIT
Conflicts: ggml, ik_llama.cpp, libggml, llama.cpp
Provides: ik_llama.cpp
Submitter: ryant0000
Maintainer: ryant0000
Last Packager: ryant0000
Votes: 0
Popularity: 0.000000
First Submitted: 2026-02-15 18:58 (UTC)
Last Updated: 2026-03-05 23:48 (UTC)

Dependencies (14)

Required by (0)

Sources (3)

Latest Comments

shaybox commented on 2026-06-25 23:11 (UTC)

/home/.../src/ik_llama.cpp/examples/rpc/rpc-server.cpp:266:37: error: too few arguments to function ‘ggml_backend* ggml_backend_cuda_init(int, const void*, const void*)’ 266 | backend = ggml_backend_cuda_init(device, nullptr); // init device

ryant0000 commented on 2026-05-29 18:01 (UTC)

I was under the impression the MAKEFLAGS variable in /etc/makepkg.conf should cover that, is that not correct? I can confirm on my system it builds following that value.

NoelJacob commented on 2026-05-26 19:34 (UTC)

Add -config Release -- -j $(nproc) to build

ryant0000 commented on 2026-03-05 23:49 (UTC)

Added nccl as dependency

ItachiSan commented on 2026-03-05 16:41 (UTC)

I noticed in the log that nccl is looked up. I would recommend adding it as dependency.